Nagra D open-reel digital recorder Page 3
My impressions of the Nagra-D were gleaned on three separate occasions: in November 1993, when I recorded Canadian pianist Robert Silverman performing the Liszt Piano Sonata in b and other works by Liszt; and in July and October '95, when I recorded, then edited, five days and nights of concerts and rehearsals at the 1995 Santa Fe Chamber Music Festival to produce our Festival CD.
Operationally, the unit was a dream. Once I had found my way around the menu (and discovered the essential Escape key combination), used the menu to set the tape tension to that appropriate for the Ampex 467 tape I was using (40 grams; BASF 931 is the same, while 3M/Imation tape requires the alternate 50gm setting), and set the four inputs to "analog/line" for the Silverman (where I used an external Sonosax 4-channel mike preamplifier), or to "analog/microphone/48V phantom power" for Festival, there was nothing more to do other than set the input sensitivities, set the EQ to Flat, make sure the monitor switch was set to "Auto" (mike feeds with the tape stationary; off-tape in Record or Play), turn the meter illumination on, and feed the machine 5" reels of digital tape at $18 each as required.
And the joy of on-location recording where, headphones, microphones, cables, and stands apart, the entire system—digital recorder/mike preamps/ADCs—could be picked up with one hand and carried out to the car at the end (footnote 5). In fact, when Audiofon's Peter McGrath recorded all the live music for us at HI-FI '95, Stereophile's Los Angeles high-end audio show, last April, I saw him walking in with his Nagra-D in one hand and a case with his Schoeps KFM-6 Sphere stereo microphone and cables in the other—the ultimate portable recording rig.
But portability and practicality aside, the point of using a Nagra-D is sound quality. How does it sound? Comparisons of 20-bit data and that same data truncated to 16 bits, using the Sonic Solutions hard-disk editing system, have convinced me that music recorded with more-than-16-bit resolution is qualitatively different from the usual sound of CDs, so the real question is: How do the Nagra-D's A/D converters sound?
The only comparable ADC I had was the UltraAnalog-based 20-bit Manley (footnote 6), which Stereophile had purchased to make the CD masters for our first three recording projects, Poem, Intermezzo, and Concert, and which had been the overall winner at the 1991 ADC "shootout" at the New York AES Convention. In comparisons with the mike feed, the Manley and Nagra ADCs did sound different, perhaps partly due to differences in the analog signal path ahead of the oversampling ADC chips themselves. Despite being all solid-state, the Manley had more a tubelike warmth to its sound. By contrast, the Nagra-D's ADCs were more cool-sounding. Both, however, had similarly impressive abilities to produce two-channel recordings that threw an immense stereo soundstage when played back via the Mark Levinson No.30.5 DAC in my usual reference system. Resolution of fine detail, reverberation tails, and the like didn't appear to significantly differ.
Given my druthers, I preferred the Manley ADC, given that a warm balance only tends to get cooler-sounding through the vagaries of CD production. But regarding overall quality, I have to say that there was very little in it. And given that I currently like to use arrays of four microphones for my recordings—a central pair for precise imaging plus spaced B&K omnis to add low-frequency bloom—the fact that I only had two channels of Manley A/D but four channels of Nagra A/D became the deciding factor. There was also the small matter that whatever converters are used, they all have to be synched to the same word clock, something that is trivial using the D's internal ADCs (footnote 7).
I suspect that my throwaway comment on the subjective quality of 20-bit digital sound has left you wanting more. Well, the differences between 20-bit and 16-bit renditions of the same music proved fascinating because it was in the solidity of the bass that I felt the extra four bits proved their worth—that, and in the feeling of U-R-There realism. Soundsources recorded on the Nagra-D sounded simply more palpable—and nothing like CD! There was simply a greater sense of hush between the notes compared with 16-bit. And as I say elsewhere in this issue, 20-bit applause sounds likes hands clapping; 16-bit applause sounds like applause-flavored noise. While it can be an acceptably good transfer medium, the 16-bit DAT doesn't even begin to compare with the Nagra-D in ultimate sound quality.
Does the 20-bit Nagra produce recordings better than the highest-quality analog recorders? It should be noted that I have no experience with other high-end 20-bit digital recorders, like the Sony DASH machines. However, over the past two decades, I have used 3M, Ampex, Otari, ReVox, Studer, Tandberg, Tascam, and TEAC open-reel analog machines, in flavors ranging from 2-track to 24-track. The tiny Nagra-D puts even the mighty ½"-tape, 2-track Ampex ATR-100 in the shade.
Yes, the Nagra-D is expensive. But without a doubt it's the finest audio recorder I've ever had the privilege to use. And the most flexible. Highly, I repeat highly, recommended.
Footnote 5: Nagra-owning recordists have their own newsletter, the Gazette and Digest for the Absolute Nagrist. Edited by Albert Swanson, GDAN is available from Location Recording, 4202 N.E. 105th St., Seattle, WA 98125. Tel: (206) 527-0318. Fax: (206) 727-6990.—John Atkinson
Footnote 6: It's a trivial matter to set the Nagra-D to accept digital inputs and lock on to a external clock. "Hot" ADCs in the professional scene are the Apogee, dCS, and Prism models, with all three manufacturers offering a 24-bit ADC! We'll be reporting next month on a recent recording session where engineer Gabe Wiener used the Prism AD-1 converter to feed 24-bit data both to a Nagra-D and the new Genex 24-bit magneto-optical disc recorder.—John Atkinson
Footnote 7: Because the same clock is used for all four ADCs, when I later uploaded the data to my Sonic Solutions editing system, I would be able to move tracks backward or forward in time with respect to the others, knowing that the time relationship would be preserved throughout the sound file. If the same word clock hadn't been used for the four channels, it wouldn't have been possible to play them back simultaneously while retaining time synchronization.—John Atkinson